Assimilation of Chemical Data from NASA s EOS Instruments: Experiences from the GMAO Thoughts for planning

Size: px

Start display at page:

Download "Assimilation of Chemical Data from NASA s EOS Instruments: Experiences from the GMAO Thoughts for planning"

Luke Powers
6 years ago
Views:

1 Assimilation of Chemical Data from NASA s EOS Instruments: Experiences from the GMAO Thoughts for planning Steven Pawson, Ivanka Stajner, Andrew Tangborn, Krzysztof Wargan of NASA s Global Modeling and Assimilation Office Daniel Jacob, Solène Turquety and the Harvard University Intex Team Kevin Bowman and the TES Team Nathaniel Livesey, Lucien Froidevaux and the MLS Team Community Workshop on Air Quality Remote Sensing from Space: Defining an Optimum Observing Strategy National Center for Atmospheric Research, Boulder, CO February 21-23, 2006

2 Objectives of Presentation To illustrate some impacts of different types of satellite data on assimilation of ozone and carbon monoxide (OSEs) To promote discussion about the role of OSSEs in future mission design

3 Impact of Aura OMI and MLS ozone data: realistic profile shape in lower stratosphere because of information content of MLS Sunrise near 37 S Sunset near 67 N Comparisons at SAGE III measurement locations on January 11-13, Mean profiles from: - SAGE III - OMI+MLS assimilation - SBUV/2 assimilation Similar results obtained by assimilating MIPAS (limb IR) and POAM/ILAS (limb occultation) data for different years and locations

4 OMI data have a beneficial impact on total tropospheric ozone columns from assimilation: Spatial and temporal variability are captured Hohenpeissenberg (47.5ºN, 11.0ºE) Lauder (45.0ºS, 169.6ºE) - Sondes - MLS+OMI assimilation - MLS+SBUV assimilation January February January February Jokioinen (60.6ºN, 23.5ºE) Marambio (64.2ºS, 235.5ºE) January February January February

kernel TES a priori TES retrieval Analyzed ozone at 150 hpa A dynamically driven low-ozone feature over the Atlantic on Feb.

5 Towards an OMI+MLS+TES assimilation: Potential capabilities of the model to isolate layering not evident with smooth nadir data mpa Forecast Forecast with aver. kernel TES a priori TES retrieval Analyzed ozone at 150 hpa A dynamically driven low-ozone feature over the Atlantic on Feb. 17 present in the assimilated (OMI+MLS) field is smoothed out in the TES retrieval. This shows that OMI+MLS+TES assimilation has the potential to resolve vertical features that are obscured in the TES retrievals.

6 Summary of EOS-Aura ozone assimilation OMI+MLS assimilation yields accurate representations of stratospheric profile and (generally) tropospheric column Optimism that TES retrievals can be incorporated, with appropriate weight given to information content (a priori, averaging kernel) Concept of OSEs (Observation Sensitivity Experiments) - combining several data types Basis for other constituent assimilation (e.g., carbon species) in GMAO

7 but, how well can present tropospheric data be utilized for Air Quality? Need near-surface analyses and (say) 24-hour forecasts, yet most present space-based constituent data constrain upper-middle troposphere - address pollution export issues, with less immediate relevance to air quality (aerosols somewhat different) Source estimates determined from space-based instruments can play an important role, as demonstrated clearly in summer 2004 (Intex-NA): Alaskan/Canadian emissions calculated from MODIS fire observations included in GEOS-4 helped link this burning to degraded air quality in Houston (Morris et al., JGR submitted) Layering of air masses exposed to various sources of pollution transported over the Eastern US

8 Global contributions to East-Coast USA pollution modeled during Intex-NA Analyzed ozone at 150 hpa Figure produced by Harvard group as part of Intex-NA mission planning

9 The following questions arise (maybe more): What is the value of nadir IR (MOPITT CO, AIRS/IASI CO/ozone) measurements for Air Quality forecasting? How adequately do space-based estimates of biomassburning emissions constrain pollutant concentrations? Are we exploiting present-day satellite observations (e.g., emissions, precursors)? What do the demands of AQ forecasting really imply for observational requirements? What are the limitations of the models used in assimilation and prediction?

10 What might be wrong with the models (specifically, the transport)? MOPITT averaging kernels Integrity of the meteorological analysis impacts the value of the assimilation and forecast Model error is much more complex in the troposphere than in the stratosphere: Have some confidence in large-scale advection Aware of uncertainties in sub-grid processes (e.g., existence of clouds) and associated transport (e.g., outflow), but do not possess a full understanding of these issues What type of observations will help us correct for detrainment occurring at the wrong level? Value of cloud observations?

11 The importance of OSSEs (Observing System Simulation Experiments) A planning tool - what impact will a new data type really have on our ability to forecast air quality? Typical application: instrument X will be launched in orbit A - what impact will the data have? Better: what would be the optimum orbit for instrument X? Best: what is the observational requirement for Air Quality and can any combination of proposed instruments meet this? Should be wary of model limitations on interpretation OSSEs require much infrastructure and planning, but potentially help the community and funding agencies make wise choices

Data Assimilation for Tropospheric CO. Avelino F. Arellano, Jr. Atmospheric Chemistry Division National Center for Atmospheric Research

Data Assimilation for Tropospheric CO Avelino F. Arellano, Jr. Atmospheric Chemistry Division National Center for Atmospheric Research Caveat: Illustrative rather than quantitative, applied rather than